A cell's behavior is governed by a complex dynamical system of genetic interactions. A central role in the understanding of the nature of living systems, their stability in a changing environment, and how such systems fail in disease, such as cancer, is played by the process of differentiation. The goal of this project is to understand this process along with cellular homeostatic stability from a systems perspective. The 'state-space' of such complex nonlinear dynamical systems, representing genetic regulatory networks, consists of all possible combinations of gene activities. The regulatory interactions result in a dynamical 'flow' in this state-space. That flow or trajectory typically reaches a recurrent pattern of activities, which constitutes an attractor or the steady-state behavior of the system. Many different trajectories typically flow to the same attractor and constitute its basin of attraction. One objective of this study is to test the hypothesis that the attractors of such networks constitute the cell types of an organism, while differentiation is precisely a route (gene expression program) from one attractor into the basin of attraction of another attractor and subsequent flow to that new attractor. Another objective is to test the hypothesis that there are several distinct paths in the state-space along which cells proceed towards differentiation. A related goal is to characterize a particular differentiation process at the gene expression level. The first specific aim - mapping the molecular paths by gene expression profiling for differentiation pathways - is intended to achieve these objectives. Finally, another objective is to study the process of cellular homeostasis on the gene expression level. The particular questions related to this objective are: do the cells exhibit homeostasis on the expression level by returning to their original states in the state-space and if so, do they retrace the same trajectory on their way back? Thus, the second specific aim - the study of homeostatic stability on the gene expression level - is proposed to realize this objective. The methods designed to achieve these goals include treating HL60 promyelocytic leukemia cells, a well-established differentiation model, with different doses and durations of all-trans retinoic acid (ATRA) and dimethyl sulfoxide (DMSO), to differentiate the cells into monocytes and granulocytes, respectively. Using early differentiation cell surface markers (CD11b) and flow cytometry, the investigators will construct loci on the dose-duration plane such that a given locus corresponds to a fixed percentage of differentiated cells. Given several different treatments that place the cells on the same locus, the cells will be profiled at different time points with microarrays in order to determine whether they follow distinct paths of differentiation. With additional microarray profiling of untreated cells, gene sets important for monocytic and granulocytic differentiation on different loci will be revealed. In order to study homeostatic stability, cells will be treated such that they are on the 50% locus and microarray profiling will be performed at different time points during treatment. After live sorting of the cells using CD11b, the CD11b positive and negative cells will be cultured in the absence of differentiation inducing agents. Microarrays will be used to profile each of these cell populations using time-point measurements, thus making possible the characterization of homeostatic behavior on the gene expression level.